Image processing device, inspection device, image processing method, and non-transitory recording medium

Abstract

An image processing device includes an acquisitor and a processor. The acquisitor acquires information including a first image of an object and a first relationship. The first image includes first and second image areas. The first image area corresponds to a first position of a first part of the object and has a first pixel value. The second image area corresponds to a second position of a second part of the object and has a second pixel value. The processor performs generating a first generated image including a third and a fourth image areas corresponding to the first and the second positions, determining a third pixel value of the third image area and a fourth pixel value of the fourth image area, and calculating a first distance corresponding to a length of the third image area and a second distance corresponding to a length of the fourth image area.

What is claimed is: 
     
       1. An inspection device comprising:
 an input/output (I/O) interface configured to acquire information including
 a first taken image including a first image of an object disposed in a real space, the first image including a first image area and a second image area, the first image area corresponding to a first position of a first part of the object and having a first pixel value, the second image area corresponding to a second position of a second part of the object and having a second pixel value, and 
 a first relationship between a position of the first image area and the first position; and 
 
 a processor configured to perform
 first image generation processing to generate a first generated image including a third image area corresponding to the first position and a fourth image area corresponding to the second position, and projecting the object on a first plane, to determine a third pixel value of the third image area based on the information and the first pixel value, and to determine a fourth pixel value of the fourth image area based on the information and the second pixel value, 
 first distance calculation processing to calculate a first distance in the real space corresponding to a length of the third image area in a first direction and a second distance in the real space corresponding to a length of the fourth image area in the first direction, the length of the third image area in the first direction being equal to the length of the fourth image area in the first direction, and the first distance being different from the second distance, and 
 inspection processing to inspect abnormity of the first part of the object when the first distance is one of equal to and shorter than a predetermined value. 
 
 
     
     
       2. The device according to  claim 1 , wherein
 the first image area is an area of one of pixels of the first taken image, 
 the second image area is an area of another of the pixels of the first taken image, 
 the third image area is an area of one of pixels of the first generated image, and 
 the fourth image area is an area of another of the pixels of the first generated image. 
 
     
     
       3. The device according to  claim 1 , wherein
 the information related to the first relationship includes imaging parameters of a camera having imaged the object, the imaging parameters including a focal distance of the camera and an image center position of the camera, and 
 the first position is a position obtained by performing a coordinate conversion on the position of the first image area based on the imaging parameters. 
 
     
     
       4. The device according to  claim 1 , wherein
 the processor is further configured to perform image compression processing to compress the first generated image, and 
 the image compression processing includes setting a compression ratio in the third image area to be higher than a compression ratio in the fourth image area, when the first distance is longer than the second distance, and setting the compression ratio in the third image area to be lower than the compression ratio in the fourth image area, when the first distance is shorter than the second distance. 
 
     
     
       5. The device according to  claim 1 , wherein
 the information further includes
 a second taken image including a second image of the object and being different from the first taken image, the second image including a fifth image area and a sixth image area, the fifth image area corresponding to a third position of a third part of the object and having a fifth pixel value, the sixth image area corresponding to a fourth position of a fourth part of the object and having a sixth pixel value, and 
 a third relationship between a position of the fifth image area and the third position, and 
 
 the processor is further configured to perform
 second image generation processing to generate a second generated image including a seventh image area corresponding to the third position and an eighth image area corresponding to the fourth position, and projecting the object on a second plane, to determine a seventh pixel value of the seventh image area based on the information and the fifth pixel value of the fifth image area, and to determine an eighth pixel value of the eighth image area based on the information and the sixth pixel value, 
 second distance calculation processing to calculate a third distance in the real space corresponding to a length of the seventh image area in the first direction and a fourth distance in the real space corresponding to a length of the eighth image area in the first direction, the length of the seventh image area in the first direction being equal to the length of the eighth image area in the first direction, and the third distance being different from the fourth distance, 
 motion vector calculation processing to detect a ninth image area corresponding to the third image area in the second generated image and having a ninth pixel value, and to calculate a motion vector including a difference between a position of the third image area and a position of the ninth image area, a length of the ninth image area in the first direction corresponding to a fifth distance in the real space, and 
 third image generation processing to generate a third generated image including a tenth image area corresponding to the first position, and projecting the object on the second plane, based on the first generated image, the second generated image, and the motion vector, and 
 
 the processor is further configured to set
 a pixel value in the tenth image area to the third pixel value, when the first distance is shorter than the fifth distance, and 
 the pixel value in the tenth image area to the ninth pixel value, when the first distance is one of equal to and longer than the fifth distance. 
 
 
     
     
       6. The device according to  claim 1 , wherein
 the inspection processing includes detecting the abnormity by performing a pattern analysis based on pixel values in an area including the third image area, and 
 the abnormity includes at least either one of a crack, a track, a defect, water leakage, exfoliation, and rust of the first part. 
 
     
     
       7. An image processing method, comprising:
 acquiring information including
 a first taken image including a first image of an object disposed in a real space, the first image including a first image area and a second image area, the first image area corresponding to a first position of a first part of the object and having a first pixel value, the second image area corresponding to a second position of a second part of the object and having a second pixel value, and 
 a first relationship between a position of the first image area and the first position; 
 
 generating a first generated image including a third image area corresponding to the first position and a fourth image area corresponding to the second position, and projecting the object on a first plane, determining a third pixel value of the third image area based on the information and the first pixel value, and determining a fourth pixel value of the fourth image area based on the information and the second pixel value; 
 calculating a first distance in the real space corresponding to a length of the third image area in a first direction and a second distance in the real space corresponding to a length of the fourth image area in the first direction, the length of the third image area in the first direction being equal to the length of the fourth image area in the first direction, and the first distance being different from the second distance; and 
 inspecting abnormity of the first part of the object when the first distance is one of equal to and shorter than a predetermined value. 
 
     
     
       8. The method according to  claim 7 , wherein
 the first image area is an area of one of pixels of the first taken image, 
 the second image area is an area of another of the pixels of the first taken image, 
 the third image area is an area of one of pixels of the first generated image, and 
 the fourth image area is an area of another of the pixels of the first generated image. 
 
     
     
       9. The method according to  claim 7 , wherein
 the information related to the first relationship includes imaging parameters of a camera having imaged the object, the imaging parameters including a focal distance of the camera and an image center position of the camera, and 
 the first position is a position obtained by performing a coordinate conversion on the position of the first image area based on the imaging parameters. 
 
     
     
       10. The method according to  claim 7 , further comprising
 setting a compression ratio in the third image area to be higher than a compression ratio in the fourth image area, when the first distance is longer than the second distance, and setting the compression ratio in the third image area to be lower than the compression ratio in the fourth image area, when the first distance is shorter than the second distance. 
 
     
     
       11. The method according to  claim 7 , wherein
 the information further includes 
 a second taken image including a second image of the object and being different from the first taken image, the second image including a fifth image area and a sixth image area, the fifth image area corresponding to a third position of a third part of the object and having a fifth pixel value, the sixth image area corresponding to a fourth position of a fourth part of the object and having a sixth pixel value, and 
 a third relationship between a position of the fifth image area and the third position, and 
 the method further comprises
 generating a second generated image including a seventh image area corresponding to the third position and an eighth image area corresponding to the fourth position, and projecting the object on a second plane, and determining a seventh pixel value of the seventh image area based on the information and the fifth pixel value of the fifth image area, and determining an eighth pixel value of the eighth image area based on the information and the sixth pixel value, 
 calculating a third distance in the real space corresponding to a length of the seventh image area in the first direction and a fourth distance in the real space corresponding to a length of the eighth image area in the first direction, the length of the seventh image area in the first direction being equal to the length of the eighth image area in the first direction, and the third distance being different from the fourth distance, 
 detecting a ninth image area corresponding to the third image area in the second generated image and having a ninth pixel value, and calculating a motion vector including a difference between a position of the third image area and a position of the ninth image area, a length of the ninth image area in the first direction corresponding to a fifth distance in the real space, and 
 generating a third generated image including a tenth image area corresponding to the first position, and projecting the object on the second plane, based on the first generated image, the second generated image, and the motion vector, and 
 setting a pixel value in the tenth image area to the third pixel value, when the first distance is shorter than the fifth distance, and 
 setting the pixel value in the tenth image area to the ninth pixel value, when the first distance is one of equal to and longer than the fifth distance. 
 
 
     
     
       12. The method according to  claim 7 , wherein
 the abnormity inspecting comprises performing a pattern analysis based on pixel values in an area including the third image area, and 
 the abnormity includes at least either one of a crack, a track, a defect, water leakage, exfoliation, and rust of the first part.